Problem statement

Issue that we are trying to address

Cold war, that spanned from 1947 to 1991 was one of the most trying phases in the world history. The two sides of the world, the Capitalist Western bloc comprising of USA and its NATO allies and Marxist-Leninist Eastern bloc led by Soviet Union (USSR) and its satellite states were at loggerheads with each other. Aside from the development of the two sides’ nuclear arsenals, and their deployment of conventional military forces, the struggle for dominance was expressed via proxy wars around the globe, psychological warfare and massive propaganda campaigns[1].

Post-cold war hopes of a world without wars, where conflicts between and within nations are peacefully resolved through negotiations, have not been fully realized. Recent data on armed conflicts indicate that during the period 1989-2016, there were 997 such conflicts in 117 countries around the world. [2].

The main purpose of the analysis conducted in the report is to have a deeper and clearer understanding of the conflict zones throughout the world, the countries involved in direct conflicts, the countries involved in proxy wars and how has such a volatile scenario shaped the way the countries interact with each other in the present day world. This helps us understand the political and social climate in the countries that have transformed themselves from a Fragile state [involved in inter-state and intra-state conflicts] into a Stable state and those which end up becoming a Failed state.[3]

Identifying Target population, \({P}_{Target}\)

The population of interest about which we want to learn some attributes is called target population.

Unit

The basic unit of analysis in the Uppsala Conflict Data Program-Georeferenced Event Dataset [UCDP-GED] Version 17.1 is the “event”, i.e. “An incident where armed force was used by an organised actor against another organized actor, or against civilians, resulting in at least 1 direct death at a specific location and a specific date”.[4]

The target population will also include the set the events that have not been recorded in the the UCDP dataset either because they were not reported or they were not clear enough from the sources to have occured. Also, any ongoing conflicts or conflicts that might happen in future form an essential component of the target population.

Variates in the dataset

##  [1] "id"                "year"              "active_year"      
##  [4] "type_of_violence"  "conflict_new_id"   "conflict_name"    
##  [7] "dyad_new_id"       "dyad_name"         "side_a_new_id"    
## [10] "gwnoa"             "side_a"            "side_b_new_id"    
## [13] "gwnob"             "side_b"            "number_of_sources"
## [16] "source_article"    "source_office"     "source_date"      
## [19] "source_headline"   "source_original"   "where_prec"       
## [22] "where_coordinates" "adm_1"             "adm_2"            
## [25] "latitude"          "longitude"         "geom_wkt"         
## [28] "priogrid_gid"      "country"           "country_id"       
## [31] "code"              "region"            "event_clarity"    
## [34] "date_prec"         "date_start"        "date_end"         
## [37] "deaths_a"          "deaths_b"          "deaths_civilians" 
## [40] "deaths_unknown"    "best"              "low"              
## [43] "high"              "durationEvent"     "groupsInvolved"

Detailed description of the fields is present in the attached codebook. Other datasets will be explained as and when required.

Understanding the dataset

Source of data

Main Data sources:

ArmedConflict.csv
ExternalHelpLarge.csv

Auxillary Data sources:

TotalWorldPop.csv
CountryLatLong.csv

The original reporting underlying UCDP GED[4] are collected from three sets of sources:

global newswire reporting
global monitoring and translation of local news performed by the BBC
secondary sources such as local media, NGO and IGO reports, field reports, books etc.

This data collection is done via a two pass process. The first step involves consulting newswire sources from around the globe. The second step involves consulting local/ specialized sources based on the data collected from the first pass.

Sources for first pass:

Reuters News,
Agence France Presse
Associated Press
Xinhua
BBC Monitoring

Sources for second pass:

local monitoring of various local media
local monitoring and research organizations
global NGO reports
UN, EU, AU and other IGO reports
governmental publications
research articles or books etc

The decision to whether or not the secondary source from second pass should be included is based on the conclusion reached at the end of first pass. If the infirmation gained after first pass is sufficient, then secondary source is not included.

Adding new columns to the dataset

I have added multiple columns to the existing data.

Dataset: ArmedConflict.csv

code: ISO Country code
durationEvent: Event duration in days.
groupsInvolved: Groups involved in conflict.

Dataset: ExternalHelpLarge.csv

code1: ISO code for country on side1
country1_lat: Latitude of country on side1
country1_long: Longitude of country on side1
code2: ISO code for country on side2, ‘NA’ otherwise
country2_lat: Latitude of country on side2
country2_long: Longitude of country on side2
external_code: ISO code for country giving external support.
external_lat: Latitude of country providing external support
external_long: Longitude of country providing external support

Data validation and reliability

Difference in reporting of various events has an impact on the level of detail to which the information can be extracted and hence included in the dataset. For example, reporting of precise location of the occurence of event might be uncommon and hence not possible to capture in the dataset at hand.

Assurance of quality of the dataset

Multiple steps have been taken to ensure quality of the dataset:

Two stage coding procedure which involves atleast two different coders being in charge of coding and revising the events before the final dataset is released. This is done atleast one year apart so that the two coders are not influenced by each other and hence ensure inter-coder reliability.
Pre-defined set of protocols to be followed by coders at all times to ensure that highest quality product is delivered.
Inconsistencies in the data are resolved by extensive background research combined with frequent meetings.
Around 50 automatic checks followed by manual checking by person in-charge is done on the dataset before it is finally made public.
Any change to be made to dataset involves some kind of human intervention to ensure that the edge cases are not missed or considered as outliers.

Pattern exploration in the dataset

1. Distribution of conflicts based on region:

As is evident from the above barplot, Africa has seen the maximum number of armed conflicts [592] in the world from 1989 to 2016. Curiously, Middle East has much lower number of armed conflicts [57] because:

The war against Islamic State [IS] in Syria and Iraq is not included in the dataset because it is very difficult to keep track record of the conflicts occuring in the afflicted countries.
The plot captures UNIQUE conflicts. In the Middle East, the same conflict has been going on and on for years together and thus, it is included only once in the above plot.

2. Trend of the conflicts happening throughout the world:

## Time Series:
## Start = 1989 
## End = 2016 
## Frequency = 1 
##  [1] 37 25 22 32 39 30 19 28 29 29 39 35 33 38 44 37 41 22 32 50 29 24 36
## [24] 54 81 45 42 25

Analysis of the above graphs reveal that the year 2013 saw a jump from 54 armed conflicts to 81 unique armed conflicts. This included the all-out attack to destroy Taleban, Al-Qaeda and its affiliates spread in different countries in West Asia and Middle East. Hence many wars were waged simultaneously to rip the aforementioned terrorist organizations.

3. Region wise trend of armed conflicts:

As can be clearly seen from the line plot above, there has been a general upward trend in the conflict events in Asia with a huge increase in years spanning from 2001 to 2008. This can be explained due to the highly unstable situation in the Asian country of Afghanistan. The country has been under constant attack by Taleban and Al-Qaeda forces and is notorious for providing safe haven to many other terrorist organizations.

The country of Pakistan has also been under constant terrorist attacks in the recent years with many organizations like Tehrik-e-Taliban Pakistan [TPP] providing active support to Taleban and Al-Qaeda.

Also, hostility between the governments of India and Pakistan has been extremely high with both the sides exchanging fire on a regular basis. The situation is dire particularly in the controversial area of Jammu and Kashmir, with Kashmiri insurgents regularly attacking the civilians and army outposts.

All the above factors combined have lead to such a huge jump in the armed conflict events in Asia in the recent years.

Africa has also seen some of the deadliest conflicts in recent times with majority of the constituent states indulged in civil conflict. Such a volatile environment in which there is always a power struggle between different factions provides a breeding ground for extremist organizations to establish themselves e.g. Boko Haram in Nigeria and Salafist Group for Preaching and Combat [GSPC] in Algeria which formed basis for al-Qaeda in Islamic Maghreb [AQIM]. From 2007 onwards, Operation Juniper Shield, a joint operation conducted by USA and partner nations in the Sahara/Sahel region of Africa, consisting of counterterrorism efforts and policing of arms and drug trafficking across central Africa[5], has led to increased engagement between the governemnt forces and terrorist otfits in the African region.

Also, as expected the region of Middle East has been in a constant state of war with many regional and international powers playing proxy wars in the afflicted areas of Syria, Iraq, Yemen, Gaza etc. The dramatic increase started in 2003 with the US invasion of Iraq [6], worsened by the start of Arab spring in 2010 in the Tunisia[7] and US invasion of Libya in 2013. Ever since, Middle East has not been able to arrive at a stable state.

4. Barplot to show violenceType: State, Non-state, One sided:

As is evident from the above barplot, the non state conflict is the maximum [576] as most of the countries are fighting within its borders either owing to a civil war or fight against terrorist establishments within its boundaries. The armed conflicts in which two different countries [State based] come at loggerheads with each other is pretty small[169 conflicts]. We need to remember an important scenario in which the countries don’t fight against each other directly but are indulged in proxy wars in already war-torn countries in the world e.g. Syria, Iraq, Yemen etc. Hence such engagements are included in Non state conflicts instead of state conflicts.

5. Trend of different violence types over the years [1989-2016]:

Region: Africa

Analysis for Africa:

State based armed conflict has increased over the recent years, particularly 2007 onwards. Such an upward trend can be attributed to increased attacks of terrorist outfits against the government forces. e.g. Frequent clashes between Government of Algeria and GSPC & IS, Government of Uganda and Lord’s Resistance Army[LRA], Government of Rwanda and Democratic Forces for the Liberation of Rwanda [FDLR] and so on. Next it can be attributed to territorial conflicts owing to historical reasons amongst many other. For example, Conflits between governments of Sudan and South Sudan, between governements of Ethiopia and Eritrea. Owing to the current volatile situation in African countries, it is highly unlikely that the situation will improve.
Non-state armed conflict involves non government entities at loggerheads with each other. This can primarily be due to access for territory, grazing land, exerting formal/ informal authority, access to lootable resources like livestock and most importantly ethnicity and religion. We observe significant increase in such conflicts in 1992 owing largely to frequent clashes between Supporters of African National Congress and Inkatha Freedom Party and 2014 due to clashes based on religion (Muslims and Christians).
One sided armed conflicts occur when the civilians are specifically targeted. Some of the several reasons for this might be to coerce the civilians into submission and hence have the support of the local population, weak groups can be more prone to targeting civilians as it might be one of the ways the group secures the resources that it needs to function effectively and lastly because of ethno-religious divides [8]. In the plot, we see a spike in the year 1994 which can be majorly attributed to the Hutu-Tutsi clashes in Rwanda, attacks by Armed Islamic Group (GIA) in Algeria, National Union for the Total Independence of Angola (UNITA) in Angola and so on. Also, the spike in year 2015 can be attributed to attacks by terrorist outfits like IS, LRA and others.

Region: Asia

Analysis for Asia:

State based: Throughout the years under study, state based violence has invariably increased in Asia with huge spikes seen in 2001 and onwards. This can be attributed to the ever so increasing activities of the terrorist outfits in countries like Afghanistan and Pakistan and as a result increased military intervention by home and western governments.
One sided: An increase is observed in years from 1999 to 2008 which coincides with the peak of Taliban rule in Afghanistan and hence frequent attacks against civilians. Also, it coincides with the peak of Kashmir insurgency in India and hence increased attack on civilians.

Region: Americas

Analysis for Americas:

State based armed conflict saw a spike in 2004 owing to fights between Government of Colombia and Marxist rebel groups like National Liberation Army (ELN) and Revolutionary Armed Forces of Colombia (FARC) which subsequently subsided.
Interestingly Non state based armed conflicts in Americas can attributed to armed conflicts between various drug cartels operating in Latin America especially Juarez and Sinaloa cartel. Clearly the trend has been on a rise as more and more cartels are getting involved in bloody conflicts to earn more authority in the contested areas.

Region: Europe

Analysis for Europe

It is important to keep into consideration that Russia is included as a part of Europe in the dataset. Keeping that in mind, the peak in state based conflict in 1992-1993 can be attributed to the post Cold war breakup of USSR which resulted in multiple newly independent states fighting amongst each other for territory and ethnic reasons. For example Bosnian War, Mondova versus PMR etc .

Non state conflicts have been pretty much non existent in Europe.

Region: Middle East

Analyis for Middle East:

As can be seen from the line plot above, state based conflict is on rise in the middle east. We see a sharp increase in 2003 which is attributed to all the events that are associated with US invasion of Iraq. In the recent years from 2010 onwards, we see a sharp increase in the number of conflicts associated with the events revolving around the Arab Spring in 2010. This power gap has lead to sprouting of many terrorist organizations in the region. e.g. ISIL etc.

6. Different countries involved in conflicts region wise [1989-2016]:

Region: Asia

From the above barplot, it can be seen that Afghanistan, Pakistan and India are some of the worst affected countries by armed conflicts in Asia region.

Region: Africa

From the above stacked barplot, it can be seen that Nigeria, Somalia, Sudan and Uganda are some of the worst affected countries in the African region with extremely high frequency of Non state conflict. This can be attributed to frequent clashes between non government factions fighting to assert authority over the other.

Region: Europe

Clearly from the above barplot, we can see majority of the armed conflicts have taken place in the constituent countries of erstwhile USSR e.g. Ukraine, Yugoslavia, Georgia etc.

Region: Americas

Clearly, Mexico has disproportionately high frequency of Non state conflicts in the Americas. This is due to the thriving drug trade in the area. Different drug cartels are at loggerheads with each other and hence we see such a high number of armed conflicts in this category. Some of the most notorious cartels are: Sinaloa cartel and Gulf cartel.

Region: Middle East

As it is clear in the above plot, Iraq, Lebanon, Yemen are amongst the worst affected by armed conflicts in the middle eastern region.

Now, we have an overall idea of distributon of conflicts in different regions. The main focus of the next section will be on the impact of Non Government actors on the armed conflicts that the country experiences. This is particularly useful to analyze the situations where:

Non state actors are increasingly controlling access to territory and targeting civilian populations.
Alongside State actors being the duty-bearers on human rights obligations, non-State entities also bear human right obligations, hence their study is extremely important to understand violation of human rights.[9].
What makes Violent Non State Actors[VNSA] different from Cold-War era is the complete disappearance of “patron – proxy” relationships with the states. Now, VNSA have become independent actors in global politico-military settings. [10]
VNSA are increasingly taking control over non-governed territories of the world (failed states, lawless areas etc) [10]

For the purpose of this report, we got the non government actors in conflicts by excluding those actors that have the word “Government” in them.

7. Distribution of non state actors in various regions of the world [1989-2016]:

Region: Asia

Clearly, India, Pakistan and Afghanistan have some of the highest count of non-government actors participating in armed conflicts.

Now let’s analyze the number of armed conflicts associated with such actors in Asia.

Clearly, from the scatterplot above, there exists a postive correlation between conflict count and the number of non state actors working in a country.

Region: Africa

Clearly Nigeria, Somalia, Sudan and Democratic Republic of Congo are amongst the worst impacted by non government actors.

Now let’s analyze the number of armed conflicts associated with such actors in Africa.

Region: Americas

Mexico has the highest number of non gvernement actors involved in armed conflicts in the Americas. This is due to the large number of drug cartels functioning in Mexico.

Now let’s analyze the number of armed conflicts associated with such actors in Americas:

Region: Europe

Russia (Soviet Union) has the hihest number of non state actors involved in armed conflicts in Europe.

Now let’s analyze the number of armed conflicts associated with such actors in Europe:

Region: Middle East

Yemen, Lebanon and Iraq have the highest number of non state actors.

Now let’s analyze the number of armed conflicts associated with such actors in Middle East:

8. In-detail analysis of events making up conflicts in different regions. [1989-2016]:

Till now the analysis was done at conflict level. Now we will channel our analysis on study of individual events that make up the conflicts in different regions of the world. We will also try and find out interesting cases of conflicts worldwide and drill down deep further for analysis.

For each region, we will break down the conflicts into civilian [type_of_violence = 3] and non-civilian conflicts [type_of_violence != 3].

Region: Asia

1. Attacks not focused on civilians.

It is expected that the longest running conflicts will also correspond to maximum number of casualities on both sides of the conflicting groups along with civilian deaths as collateral damage.

Clearly, the conflict in Afghanistan has been running for 28 years. It has the maximum number of casualities [161413] associated with it. It has resulted in 12539 non combatant casualties. Sadly, the war in Afghanistan is not yet over and the country is in the constant state of turmoil. Hence we can expect the deaths to increase in near future.

On the other hand, Kashmir insurgency in India has been running for 28 years [from the start year of dataset]. But the number of casualities [17345] is far less than that in Afghanistan. Also, there are far less reported non-combatant casualities [794].

This encourages us to understand the political and military climate in the two countries that has resulted in such a huge difference in loss of life and property. The country of Afghanistan has been under attack ever since the Saur revolution of 1978. Soviet invasion of Afghanistan in 1979 continued till 1989. This led to the spawning anti-communist outfits or Mujahideen in various parts of Afghanistan. This led to the creation of one of the most notorious extremist organizations - Taleban which ultimately took control over the country from 1992-2001. In 2001, USA and its NATO allies invaded Afghanistan to eliminate al-Qaeda and Taliban. Hence the country has always been under attack.

On the other hand, the Kashmir insurgency in India is limited to the Kashmir region of India and does not engulf the entire country. Hence we can see the huge difference in the 2 countries under consideration. This can be shown by the following scatterplots and maps:

## [1] Afghanistan Pakistan   
## 117 Levels: Afghanistan Albania Algeria Angola Argentina ... Zimbabwe (Rhodesia)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   0.000   4.000   6.000   7.802   9.000  36.000

##     Min.  1st Qu.   Median     Mean  3rd Qu.     Max. 
##    0.000    1.000    3.000    7.691    6.000 2150.000

Clearly the distribution of death count for Afghanistan is highly skewed, Hence we will consider only those events that have led to greater than or equal to 25 deaths.

This scatterplot helps us understand the intensity of attacks and subsequent deaths associated with this conflict in Afghanistan.

This scatterplot clarifies the intensity of attacks and subsequent attacks in India.

It is clear from the above maps, that the conflict in Afghanistan[Taleban and affiliated organizations] has engulfed the entire country [especially concentrated in North and North-Eastern Afghanistan] whereas the Kashmir insurgency in India is restricted to the bordering areas with Pakistan.

Some other interesting conflicts that should be analyzed are:

Afghanistan- Islamic State

Ever since IS emerged in Afghanistan in 2015, it has tried to expand its control over Northern Afghanistan and this has made North Afghanistan particularly unsafe not only for indigenous population but also neighbouring Central Asian countries having close ties with Moscow. Allegedly, this has brought Taliban and Russia closer as they both see a common enemy IS. [11, 12]

Naturally the terrorist organizations attack each other as is evident in the following graph:

IS-Taleban

Clearly the areas on armed conflicts are common to the regions where both the extremist organizations have stronghold [Northern Afghanistan]. It must be kept in mind that the datset includes the events from 1989-2016. Hence the data for IS in Afghanistan is available only for 2 years, 2015 and 2016.

After Sept 11, 2001 attacks in USA by al-Qaida, USA and its NATO allies launched Operation Enduring Freedom which intended to target terrorist mastermind Osama bin Laden’s al-Qaida organization, which was based in the country, as well as the extreme fundamentalist Taliban government that had ruled most of the country since 1996 and supported and protected al-Qaida.[13].

Starting with the al-Qaeda attack on USA in 2001, USA bombed Afghanistan, Pakistan and other countries to eliminate al-Qaeda and Taliban.

Another important conflict waged in the Asian region was the Sri Lankan civil war waged by Liberation Tigers of Tamil Eelam (LTTE) that demanded for a separate Tamilian state in along the eastern coast of Sri Lanka. This is evident in the following map.

2. Civilians as primary targets

For the armed conflicts that specifically target civilians, Afghanistan still has the highest number of civilian casualities [7062] over the course of 18 years. Myanmar, on the other hand has 4273 civilian deaths associated with the Burmese genocide.

Clearly, the attacks on civilians are concentrated along the eastern borders of Myanmar. Primarily ethnic conflicts, this area has been a hotbed of conflicts between Rakhine Buddhists and Rohingya Muslims which continues even today.

Region: Africa

1. Attacks not targeting civilians

As can be seen from the above scatterplot, Eritrean–Ethiopian War (1998-2000) has been one of the bloodiest wars in the African region resulting in 78429 casualities ins short span of 3 years.

This map shows Africa’s longest running civil war that led to creation of South Sudan in 2005. This war led to 48913 deaths over the span of 28 years.

2. Attacks directed at civilians

African countries have some of the most repressive governments in the world. Most of the countries have dictators e.g. Zimbabwe whereas others are divided along ethno-religious lines.

The government crackdown on civilians in Rwanda led to the killings of 511210 civilians. The most apalling of the attacks was the Hutu government backed massacre of Tutsi peopple in 1994 that killed around 300559 civilians.

Another country that was involved in a civil war for a long time was Sierra Leone [1991-2002].

The majority of the crimes were perpetrated by rebels from the Armed Forces Revolutionary Council (AFRC) and the Revolutionary United Front (RUF).

Region: Americas

1. Attacks not targeting civilians

Columbia has pretty much been involved in a civil war that began in mid-1960s and continues even today. It is a by-product of Cold War in which USA supported strong anti-communist repression in rural Colombia in the 1960s that led to liberal and communist militants to re-organize into Revolutionary Armed Forces of Columbia (FARC) among many others [14].

Peru has been engulfed in internal conflict which involves the government of Peru and some terrorist organizations such as ‘Shining Path’ (Sendero Luminoso) and Movimiento Revolucionario Túpac Amaru (MRTA), both of which support the policy of Marxism-Leninism [15].

In Mexico, most of the armed attacks have been between Juarez and Sinaloa cartel which are international drug trafficking, money laundering and organized crime syndicates [16].

2. Attacks directed at civilians

This represents the attack by Al-Qaida on American soil that led to 2753 civilian deaths within a single day.

Region: Europe

1. Attacks not targeting civilians

The map above timelines the Russo-Chechen conflict starting in 1994 and continuing till 2007. Majority of the attacks have been in the Chechen capital of Grozny as can be seen by the concentration of attacks in the map above. It led to the death of 18,176 people over the span of 14 years. It resulted in the decisive Russian victory and establishment of pro-Russian Chechen government [17], [18]. An interesting point in the above plot is an bomb blast in Doha, Qatar to assassinate Zelimkhan Yandarbiyev [19], [20].

The above plot represents the Bosnian War from 1992 to 1995 that led to internal partition of Bosmia and Herzegovina. The main belligerants were the forces of republic of Bosnia and Herzegovina and those of the self-proclaimed Bosnian Serb and Bosnian Croat entities within Bosnia and Herzegovina [21].

2. Attacks directed at civilians.

The Bosnian War saw the first case of genocide in Europe since World War II, a grim reminder of which is Srebrenica massacre of 1995 that led the the killings of 8106 Muslim Bosniaks in July 1995 [22].

Region: Middle East

1. Attacks not targeting civilians

2. Attacks directed at civilians

Islamic State (IS) was born during the US invasion of Iraq in 2003. This terrorist outfit has attacked civilians many times sice 2004 and has even spread its influence in the Levant region and hence has become Islamic State in Iraq and the Levant (ISIL). This has led to th death of atleast 14,100 civilians since 2004.

Plot that shows multiple attacks carried out by Islamic State against the states in various countries of the world. IS has become a notorious terrorist organization the world over as is evident by the attacks it has carried against the civilians.

Palestinian militant actions escalated in the Gaza strip following the overwhelming election to government of the Islamist political party Hamas in 2005 and 2006. Presently, most of the West Bank is administered by Israel though 42% of it is under varying degrees of autonomous rule by the Fatah-run Palestinian Authority. The Gaza Strip is currently under the control of Hamas. Hamas has launched multiple offensives against the civilians as is evident from the plot above [23].

The above analysis focused on some of the most violent and long-lasting conflicts the world has seen. Many of them are still prevelant today with ever-increasing number of casualities on both sides of the conflict with no end in near future.

9. In-detail analysis of overall trend of the conflicts throughout the world. [1989-2016]:

We are particularly interested in terrorist organizations.

Which are the top 20 most notorious organizations which have led to maximum number of attacks on governments worldwide?

The above barplot tells us the top 20 organizations that had the highest frequency of attacks per active year from 1989 to 2016. As it can be seen, Taliban has the highest frequency of deadly attacks against the government [1135]. Interestingy, Islamic State (IS) ranks second closely followed by Tehrik-e-Taliban, Pakistan (TTP). Taleban, which was the by-product of Soviet invasion of Afghanistan (1979), has been active ever since. This explains the increasing number of attacks by the organization.

Taliban: The Taliban government in Afghanistan was ousted by USA and its allies in 2001. This led to the beginning of the Taliban insurgency in 2002 which continues even today. This can clearly be seen with a steep increase in number of attacks against the government [24].

Islamic state (IS): The attacks by IS have been on a rise especially after the Arab spring that swept the Middle East in 2011.

Which are the top 20 most notorious organizations which have led to maximum number of attacks on civilians worldwide?

Islamic State (IS) has been the most notorious organization in recent years that has regularly attacked civilians.

Islamic State (IS) : Clearly the attack against civilians have increased over the years starting from the Arab spring of 2011.

Taleban: It has attacked the civilians with extreme brutality which continues to this day.

Which of the terrorist organizations are most widespread?

## [1] IS         al-Qaida   GSPC       Al-Shabaab FDLR       LRA       
## 1059 Levels: 1920 Revolution Brigades - IS 3R ... Zviadists

The above maps shows the different countries that the Islamic State or its affiliates have attacked. Clearly, IS has been the most violent in Iraq [in terms of deaths incurred] where it all started. The organization was successful in carrying out multiple attacks throughout the world either conducted by ISIS or inspired by it [25]. As can be seen, the terror organization has swept most of the Sunni dominated countries in the Middle East and Northern Africa [26]. The United States of America has also been attacked by ISIS-inspired organizations [27].

Al-Qaida rose up in response to the the Soviet invasion of Afghanistan in 1979 in which the locals and foreign elements [like Osama Bin Laden] came under the umbrella of Mujahideen to free Afghanistan of all external influence. Hence it has caused maximum number of deaths in the Asian countries of Afghanistan and Pakistan. The terrorist outfit marked its presence in USA with the September 11 attacks [28].

Salafist Group for Preaching and Combat (GSPC) began its attacks in 2007 in the North African country of Algeria and has spread in the surrounding countries. The terror group officially pledged support to al-Qaeda in 2007 and has since formed the basis of al-Qaeda in Islamic Maghreb (AQIM). The group has since declared its attention to attack European and American targets. It aims to overthrow the Algerian government and institute an Islamic state [29,30].

Harakat al-Shabaab al-Mujahideen is a jihadist fundamentalist group based out of Somalia in Horn of Africa. In 2012, it pledged its allegiance to al-Qaeda. It supposedly has links with the West African islamist organization of GSPC, now AQIM.

The Democratic Forces for the Liberation of Rwanda (FDLR) is an armed group active the Democratic Republic of Congo.

Lord’s Resistance Army (LRA) is a rebel and heterodox christian cult that operates out of Uganda.It has been accused of widespread violation of human rightsincluding murder, abduction, mutilation, child-sex slavery and forcing children to participate in hostilities [31].

Countries that have incurred maximum deaths owing to armed conlicts in years from 1989 to 2016

As is evident from the above map, some countries have had a long history of violent attacks like India, Afghanistan, Pakistan and so on. On the other hand, other countries like Iraq have seen the emergence of some of the world’s most notorious terror organizations like ISIS. The conflicts in these countries is pretty much ongoing with no end in near sight.

In most of the wars that are being carried out in the world today, many of the international powers are involved directly or indirectly. The reason for involvement might range from providing military support, weapons and so on to providing intelligence and logistics. One of the many reasons of involvement is that the external powers have to further their interests by involving in proxy wars. This has been a very common trend since Cold War and continues till today.

Analysis using Hypothesis Testing.

1. H0: Number of conflicts in the world is independendent of population of the world.

HA: Number of conflicts is dependent on population of the world.

Let’s plot a scatterplot to get a cursory idea.

Let’s perform hypothesis testing.

It can clearly be seen from the above plot that under null hypothesis, we have high probablility of observing something we have observed in the data. Hence we do not have evidence against the null hypothesis that the number of conflicts is independent of population of the world over the years. (p-value = 0.253)

2. H0: Number of unique state based conflicts is independent of population of the world.

HA: Number of unique state based conflicts is dependent on population of the world.

Let’s plot a scatterplot to get a cursory idea of relation.

Clearly we can see a positive correlation between the two variables. Let’s evaluate the correlation between the two.

## [1] 0.9094169

The two variables are highly correlated with corr value of 0.909.

Let’s perform hypothesis testing to determine the valididity of our hypothesis.

It can clearly be seen from the graph that the probability of observing something as weird as the data is very very low under the null hypothesis. Hence, we have evidence against the null hypothesis that number of state based conflicts is independent of population of the world.

3. H0: Number of unique non-state based conflicts is independent of population of the world.

HA: Number of unique non-state based conflicts is dependent on population of the world.

Let’s plot a scatterplot to get a cursory idea of relation.

Let’s perform hypothesis testing to determine the valididity of our hypothesis.

It can clearly be seen from the graph that the probability of observing something as weird as the data is very very low under the null hypothesis. Hence, we have evidence against the null hypothesis that number of non state based conflicts is independent of population of the world.

4. H0: Number of unique one-sided conflicts is independent of population of the world.

HA: Number of unique one-sided conflicts is dependent on population of the world.

Let’s plot a scatterplot to get a cursory idea of relation.

Let’s perform hypothesis testing to determine the valididity of our hypothesis.

It can clearly be seen from the graph that the probability of observing something as weird as the data is very very low under the null hypothesis. Hence, we have evidence against the null hypothesis that number of one-sided conflicts is independent of population of the world.

Future scope

There are multiple things that can be further analyzed with the given database.

I have not used the information regarding the source of news given by ‘source_article’ column in ArmedConflict.csv. By analyzing the information, we can try to determine the bias that a given news source has in terms of presenting the news.
We can use the information regarding the precision of location where the attack has occured using columns ‘where_prec’ and ‘where_coordinates’ in ArmedConflict.csv to determine how often a specific region is attacked in an entire country. This can help us to identify the strongholds of organizations in the country of interest.

Bibliography

Extra information [Please read]

Please find attached all the data files used for analysis in the report. To run the project place the files in a folder and replace dataDirectory ‘/Users/hardiksahi/Second Term/Stat 847/Project/FinalDataSet847’ with the location of your folder on line 61.
This report is submitted by:

HARDIK SAHI

**University of Waterloo

Armed Conflict [1989-2016]

Problem statement

Issue that we are trying to address

Identifying Target population, \({P}_{Target}\)

Unit

Variates in the dataset

Understanding the dataset

Source of data

Adding new columns to the dataset

Data validation and reliability

Assurance of quality of the dataset

Pattern exploration in the dataset

Analysis using Hypothesis Testing.

Future scope

Bibliography

Extra information [Please read]